CEOInterviews.AIFrom Meta PM to Appointing an AI CEO: A Conversation with Xiaoyin Qu

H

Host0:00

Hi guys. Hi. Welcome to Silicon and Spice where TED leaders share life's unique flavors. Today we have my longtime friend Shiain here. Do you want to wave? Hi. Hi. Hi guys. So I wrote your intro. It's pretty long. So you currently you are a founder and board chair of Hey Boss AI and previously you were the CEO of founder run the world. So it's a virtual platform that Dr. Prior in 2023. So something kind of interesting is that I got to know you actually through that platform because you organized like with the PM community like events community.

X

Xiaoyin Qu0:30

Events community through events platform. Yes. Yes.

H

Host0:33

And then prior to that you were a senior PM at Meta for three plus years and you're also a TED influencer that you really speak the truth not like fluff. And then you're also an angel investor pretty active on you know all different kind of startup communities. Yeah. And you're also mom. So it's pretty amazing there.

X

Xiaoyin Qu0:48

Yeah. Thank you. Thank you.

H

Host0:50

So my very first question is kind of about your journey from China to Silicon Valley. You and I both came to US when we were 18, right? So you studied economics at Pomona College and you also co-founded your first education startup when you were 20.

X

Xiaoyin Qu1:07

Yeah.

H

Host1:07

So I looked at the website is still up and running and still hiring and stuff.

X

Xiaoyin Qu1:11

Actually my co-founder is start running the company. It's one of the top education for college counseling. That's college counseling. Yeah. And then help people like that's a real product market fit by there.

H

Host1:23

And then you also landed in Facebook's RPM project, right? And then looking back and my question is what was the moment you realize, hey, I want to build things for my own or for others versus being a big corporation.

X

Xiaoyin Qu1:39

Well, I mean I had a great time at Facebook. I like to call it not Meta. It was a great great time was the peak. So, obviously I learned a lot, but I think I always kind of know that I wanted to start my own thing, especially when I was in college. I tried this like I try to start my foot.

H

Host1:52

You already test the water. Yes, it's very cool. But I mean, frankly speaking, probably from a like cash standpoint, it's better if you work at Anthropic 5 years ago. You probably make way more money and more reliably. But I guess I really enjoy the feeling of starting your own.

X

Xiaoyin Qu2:07

That's your passion, right? Yeah, I think so. Yeah. Yeah. I think at some point I mean maybe if I make enough money I was thinking oh I'll probably start a company anyway so it's kind of like I'm doing what I like you know I mean.

H

Host2:17

It's more that that's part of your entrepreneurial nature. I think you enjoy that right yeah.

X

Xiaoyin Qu2:22

Yeah I think so.

H

Host2:23

Okay okay.

X

Xiaoyin Qu2:24

Yeah maybe a lot of people who work in big corporate they probably want to start their own company too it's just like they maybe think oh I should make enough money and then I can do it right so so I think there's going to be a lot of people who want to do it actually yeah.

H

Host2:36

Yeah yeah no that's true that's true it kind of goes back to like nurture versus nature. But I assume when you were growing up, you had this moment you realize, hey, let me try different thing and there would be some time that you realize I want to set pace for my for my own, right?

X

Xiaoyin Qu2:50

Yeah. I think well I mean obviously I think starting company is very risky. At the same time it's pretty enjoyable. So for me it's more just like a lifestyle. I chose this lifestyle and that's quite fun.

H

Host3:02

Yeah. I remember we had this conversation years ago is that kind of just echo back what you were mentioning earlier that even if you made tons of money, you still do not want to settle to work in a company, work for others and you still want to start something.

X

Xiaoyin Qu3:15

Well, I think if you make a lot of money, probably you shouldn't deal with all the politics when you don't need to.

H

Host3:20

Right. Right.

X

Xiaoyin Qu3:21

So, I think that's obviously maybe an easier decision in that sense, right?

H

Host3:24

Yeah, sure.

X

Xiaoyin Qu3:26

Like once you have enough money, you'll do whatever you want. You work at Facebook or Google and other things, right?

H

Host3:31

Yeah. Yeah. Yeah. No, that's a good point. Kind of that kind of a segue to my next question is that about your Stanford dropout decision from the prestigious MBA program. Yeah. So, you dropped off after year one and that then to start run the world. So, that's a pretty big swing because getting into Stanford was not easy, right? Like you were surrounded by all these ultra successful with high potential all these classmates. Yeah. But then what was the moment you realized hey is there like a specific week or there specific action you realize hey I want to do something right now at this moment otherwise I'll regret.

X

Xiaoyin Qu4:04

Well not really it's just I tried many different ideas and during the summer of the first year we made the most traction from one of the idea we instantly got a lot of traction a lot make a lot of revenue actually okay so we got some term sheet from Andrew and war so it seems like a no-brainer is like you dropped out then yeah it's not I guess I always know that I may about if I find a good idea. And actually, I was not alone. There was like 12 people in my class who dropped out. So, it's not that I guess I'm People may think that's really weird, but it's really not that weird. In my class, there's like a lot of people already dropped out before me. But yeah, but but I I would just say like I feel the MBA program is very fun from a social standpoint. You meet a lot of cool people. Looking back, if I guess because I was already in tech, I don't really need MBA degree to find another tech job. So, there's not that much difference if you drop out earlier than versus graduating. But if you obviously if I work at McKinsey, I wanted to move into tech, maybe I should finish the school and then you know find a job. But really for me that back then there was not that much difference one year or the other. It was also during COVID so it was like I'm not going to be able to socialize anyways. So yeah so that seems like a very easier decision much easier decision for MBA program I guess. Yeah.

H

Host5:15

Yeah. Yeah. It's also very expensive. I mean it was very expensive.

X

Xiaoyin Qu5:18

Yeah. I hear you. So I think so sounds like it's because you found the product market fit, you draw the revenue, you got the traction, you draw the funding that makes the dropout decision much easier, right? Versus if somebody just say, "Hey, you know, this is not what I expected and then they pursue something out of their own without any safety net and that's a different story, right?" I think so. Yeah.

H

Host5:41

What were some other ideas you tried that did not work out? If you can share.

X

Xiaoyin Qu5:44

Well, I tried like some mental health idea.

H

Host5:46

Okay.

X

Xiaoyin Qu5:47

I think I I I do think that could still work. It just that exact arrangement. The exact map didn't work. Okay. I tried some other random stuff like some community app and whatever. Back then there was like a community was like very cool. Yeah. So there's a bunch of different idea that I tried. Okay. But I think that when when I host my first virtual event, that's when it really kind of got more traction and I got more conviction. Yeah. It's just like Yeah. And then I had a co-founder who also quit Facebook. So like everything kind of just like worked out, you know? Yeah. All the stars were aligned. Yeah. Exactly. Yeah. Yeah. Exactly.

H

Host6:20

But I feel like right now it's a very different world. Right now you can start a company without raising a lot of money.

X

Xiaoyin Qu6:26

Right. You can test idea much sooner than before. It's like okay tech is like four five months to build the product and see if that works or not. So we figure out a way to without building the product first test the waters to see if that works. Yeah. But now I think with AI like you can quickly build something in a week you will know right. So, so I think that the risk probably is like is much easier to know if your thing works or not in the AI world. Yeah. Because it only take you a week, right?

H

Host6:49

Right. Right. The cycle to test the water is shorter. The amount of risk you take on is also smaller, right? So you could try different things like this. Yeah. Yeah.

X

Xiaoyin Qu6:59

So the next question is regarding kind of founding Hey Boss and with your AI pivot, right? So after a successful exit or run the world, so you started Hey Boss, right? So you guys are an AI agency coding.

H

Host7:11

Web coding. Yeah. Developer agency and then it basically you can build a full website. So you started this AI company with agency. You can build websites and then you also have steel balls and then you also like before we start this podcast you showed me something you guys are brewing. So I want to hear kind of more about what made you want to go from like virtual events to like a completely different direction. Was there connecting threads or is it completely braid? you just want to start something fresh.

X

Xiaoyin Qu7:38

Well, first we sold the company 23. I would say I'm that was the generation of SAS. I basically I was in the generation of SAS. Okay. Right. So like SAS has a peak moment I think 2021 2022.

H

Host7:51

True.

X

Xiaoyin Qu7:51

Then it's like everything kind of died down a little bit. I think between 2022 2023 there didn't seem to be a lot of new things going on. ChatGPT came into play. So it's kind of like for me even like my new company now we're doing AI. So to some degree from 2023 to 2024 I kind of did a pivot from the more traditional SAS to AI right kind of I guess I probably pivot sooner than a lot of people but it's still a very different world for me so now I'm like has been like working in entrepreneurship in AI native product for two plus years I yeah I would love to kind of share about that I think a lot of people probably wanted to get into AI and I think that just just for me is like obviously there's no like if you're a 2023 2024 you're not starting a company AI then what are you doing right it's like it kind of like that's where things are you got to do something with yeah you got to do something and especially that is the biggest like wave so I guess everybody who are in tech if you're doing a startup probably should obviously probably should all do at least test it test it right now I would even say that if you don't do AI I don't think there's any funding for any company that true does not have AI right like every single company has AI but I do think that there's a lot of differences between how we run the business in the past which is in 2021 it's all about hiring more people. I have like 70 people that I managed in my company it's like oh you quickly hire a lot of people. We talk about how many people you have in your company. But now it's like the opposite. It's like oh if you have 10.

H

Host9:19

Yeah

X

Xiaoyin Qu9:19

Because like you know if you have 20 people like maybe that's okay. If you're like a billion dollar company that's okay. If you're a seed stage company and you have only 20 people like what are you doing with your life? Right? And then if you have one person, oh my god, that's like so brilliant. You know, you can do everything. Yeah. I feel like the entire, you know, the judgment of entire like I say entirely shipped, right? Success metrics of success metric. Exactly. So one is like how fewer people you have. If you have maybe five, that's okay, tolerable. More than five, like what the hell right? That's like number one, right? Number two is like, okay, it's like it seems like the company doesn't really have that many like and then it's like okay I even feel old sometimes like you know 18 year old 15 year old 16 year old and 19 year olds it's like okay like there's actually not that much benefit you have by being like old if you like have a lot of experience to some degree that experience may not directly apply hold you back exactly so to some degree like you have to I have to like at least actively unlearn what I've learned in the past it's Oh, as PM is all about alignment with other engineers, but like no, not anymore. It's like, oh, PM is all about building. So, so the whole thing I think also changed a little bit, right? So, the past experience of which, oh, I was a senior PM at Facebook, but who gives a damn you know? Like, now it's like a new world.

H

Host10:37

Yeah.

X

Xiaoyin Qu10:38

You kind of have to forget about my past experiences. And then forget about the fact that I'm like whatever have 10 plus experience. I was that doesn't matter. It's like a new world like new way of doing.

H

Host10:48

Yeah. Native. Yeah.

X

Xiaoyin Qu10:49

Yeah. Exactly. So, so I kind of have to like pretend I'm 15 again and it's like forget about all the things you learned and then try to really think for first principle. Do you really need it or not? You know what I mean? So, I think that's very different world. It's a completely new mindset.

H

Host11:03

Exactly.

X

Xiaoyin Qu11:03

Yeah. I have to entirely forget about what I the things you used to do like okay seed stage SAS also doesn't matter anymore. So even the annual recurrent revenue right now is kind of revenues like you don't really know that's true or not. And then now a lot of AI company are getting into services. They have AI employees replacing AI software. So the entire business model also changed. So I think the whole thing is like the how you make money change how you do marketing a lot of times it's like okay for SAS typically we have the product we promote it but now it's like a lot of times you tell you you sell a dream it's like oh yeah I can do this I can do that and you and there's a lot of like promotion people are spending a lot of time doing marketing like I have to do I have to be an influencer myself to do marketing is so expensive now so it's like the whole thing is like how you like sell is like okay now you have to do a lot of marketing you have to even be like Kardashian to some degree you know you have to be your own influencer.

H

Host11:54

Correct. Yeah. And that the whole thing I think the how you do marketing change, how you do product change, how you hire people change. Yeah. The whole thing, right?

X

Xiaoyin Qu12:01

So I feel like it's a very different world.

H

Host12:03

Yeah. Yeah.

X

Xiaoyin Qu12:04

I feel like everybody has to learn how to pivot into this new world. Yeah.

H

Host12:08

On the bright side.

X

Xiaoyin Qu12:09

Yeah.

H

Host12:09

If you are up and coming, you are not at a disadvantage, right? Because everybody is just starting from from zero.

X

Xiaoyin Qu12:16

And then everybody has like access to similar tools. Yeah. you're not like being gatekept from those like fancy VC dinner or some small circles because now everybody sees the same tools they can leverage there some some sort of small circles. Yeah. But but I would say like there's a lot of people who raise a lot of money but maybe there's also people who don't raise as much money they can also make good money. So yes I think that there's a lot of legacy company like you know like whatever last thing of the world you know their stock is going down.

H

Host12:46

Yeah. And so there may be a chance that you even think about oh maybe I one day I can compete against Nvidia which you will never think that's your job right even like cursor and know the whole thing changing so fast cursor was hot last year it kind of feel like every everybody at least in my opinion like all the AI company feels like a celebrity you know in those like talent show it's like this year you're hot next year next year who knows you're like out of date.

X

Xiaoyin Qu13:15

What can you tell us about the newer products that have not been released yet.

H

Host13:18

Yes or like a set set of products you guys have. Yeah.

X

Xiaoyin Qu13:21

So so we like before you know like I think the previous company Yeah. I like to believe a single product strategy. You're focusing on one product and you do really well and then you kind of keep compounding on that product. But I personally realize that if you even look at OpenAI where Anthropic they have many different products. Yep. Right. OpenAI has like you know Codex, they have like Sora I mean I guess they're shout out Sora but they have many different product and even like Jinspark for example they have the AI designer PowerPoint it's like a many product. So I feel like the whole thing is like because it's easier to create product therefore you kind of need to have many different thing going on and and also maybe you have to accept the fact that your product may be obsolete next month you got you got to keep it up. So I think the whole strategy kind of have changed and and now we have multiple product too. And the new product is we're building a platform where anybody can have their own business. We call this zero human unicorn. The idea is like you really have a team of AI agents. Yeah. That are we provide you with a team of AI agents that are well trained from coding to marketing to growth to whatever. And then we provide you with access to AI CEO that you can text and AI CEO will help you manage all the employees. And then so you can give AI CEO the idea and then the AI CEO can help you run the company.

H

Host14:36

What what are some examples you can shed little?

X

Xiaoyin Qu14:39

I know like a lot of our product skills for example is run by the AI CEO already and then so a lot of times like I can give some general ideas like for example I think oh maybe we should start filter the competitor you you should follow and then figure out if they have anything interesting.

H

Host14:54

Yeah

X

Xiaoyin Qu14:54

Just launch it. So then actually can every week they're monitor what in competitors are doing. if she find this is a good feature we should still she will just assign that to developers marketer and just launch it so that doesn't even need my involvement the whole process is just automated right or like the can you can tell them the goal right they're like okay I want to like 5x my revenue and they can figure out different ideas they want to try to to see which works which doesn't the one that doesn't work they will double down so essentially we're trying I'm we still do a little bit of a stuff but I'll say like I try to automate as much as my decision making as well as all the execution to all the AI agent And then so we still have engineer but engineer is more solving the issue where AI agents then that's what engineer are solving but they don't really have a day job as in they don't have a regular job that they're doing they don't have to do like a regular like stand up follow the spring process or a little bit yeah I guess the sprint AI is AI is doing by default AI is doing all the okay okay only when AI stuck will realize oh it's not doing a good job then the engineer's job is to teach AI how to do a better job next time but they don't actually do the job as much.

H

Host15:59

How do you evaluate that AI is doing the right thing at the right time?

X

Xiaoyin Qu16:05

Well obviously there are different ways you can go about it. You obviously can still have KPI for AI track okay and then you can make them reflect on why they are not hitting the KPI but there's obviously other things you can like taste wise you can monitor how they work and try to figure out a way to make it work. But I think because we are a smart startup and I actually made a lot of people changes in the past year when we started off when we first started Hey Boss like the product the team last beginning of last year AI coding was not a thing very so so we were still have like product we have front end we have back end we have designer we have QA we have devops now that six job become one person entirely already wow and then we're trying to be even crazier where that one person does not need to do that job. It's AI that does all the one person just monitoring okay to some degree I think it is similar to autonomous driving you have L2 L3 L4 right L5 is like entirely autonomous I think right now maybe when we first get started was like L2 we're using each of the agent is use each of the employees using some kind of to make themselves more complimentary skills yes it's like you still have the human are doing the job by default and then they're just using AI to make each person more efficient but then it's like okay it's like what can we do to merge those jobs so one person can do it all right so that case instead of six people doing one job we can do like six people doing six pod right doing six jobs so that's then you significantly increase your productivity then the next layer is like how can you make AI default do all the job and human are only there when AI is doing something stupid kind of like driving it's like AI's doing most human only intervene when things are stupid so I think that's kind of the space that we're in and I think we're eventually as AI gets better, a cloud gets better, you will eventually get into the state where you have autonomous business. I'm not saying all business can be run autonomously, but I do think a bunch of businesses can be run with significant autonomy. Yeah, you don't really need to spend like 24 like maybe eight hours a day. You can spend less. Maybe you're only unblocking AI on the things that they don't know how to do. For example, if I were to do sales, if I were to do podcast, yes, is probably you prefer to interview me versus my AI. Yeah, maybe one day AI gets so good that you don't even want to deal with me because they're funny or whatever. So, so my AI interview. Exactly. So, that hasn't happened yet, but it could. It's possible with this especially as a decision maker may be AI in the future like AI may decide what to buy. In that case, like it will be one AI interview the other AI promoting to the other AI who makes decision and human are the human is still making decision but they're making much higher level decision. Could be like I gave you a million dollar budget, you do this for me.

H

Host18:46

Yeah. Yeah. Yeah, like they don't really want to get into the day-to-day. They're more like a board member rather than like a CEO.

X

Xiaoyin Qu18:51

They're freed up to do more could be creative. Maybe you have 20 companies. Yeah. I could totally imagine a world. This is my crazy prediction. I think by the end of this year, yeah, I will have like 100 companies. And then I'll run 100 company at the same time.

H

Host19:04

Yeah.

X

Xiaoyin Qu19:05

And by the end of next year, I'll have 10,000 companies.

H

Host19:07

Yeah.

X

Xiaoyin Qu19:07

That's my genuine belief. I think that will happen.

H

Host19:10

Yeah. Yeah. Yeah. Yeah.

X

Xiaoyin Qu19:10

You know, and then I'm not saying all company will be successful. Maybe some companies succeed, some companies don't. just like the regular companies but some company only succeed for two months it's like they capitalize on a wave they make some quick bucks and then next month they're obsolete you know what I mean so so I do think that's possible I will have 10,000 companies.

H

Host19:28

So if I use analogy there there are a bunch of analogies I can think of right now so like from the car I like you when you talk about car like initially maybe you have a human driver right the human help human and then you move on to okay then you learn a skill to drive and then you move on to I'm teaching AI to drive and then you know you monitor a little bit but in the long term you give a prompt to AI then AI would drive you everywhere and then that's my first analogy that came to mind. Yeah. The second analogy it's almost when you talk about the 10,000 companies that run by let's say one or two people it's possible it's almost like you want to do like angel investment but now you're angel investing your own version of company. Correct.

X

Xiaoyin Qu20:07

Yeah. So like angel investors were YC or all these incubators like they invest in so many different diversified business.

H

Host20:14

That's a great analogy. And then now it's like boom I'm my own angel investor and then I invest in all different agents like in different business ideas right like you could stitch something see if it sticks or not to the wall and if it doesn't stick you move on to the next thing and then as you said because AI makes it easier this the switching cost from one business to another is not that bad so you could actually test out the water in faster pace and different things to do that is a pretty exciting future. Wow. I don't know but I guess I guess the thing that people are talking about is like do we really believe in like long-lasting company durable companies right because a lot of companies even like cursor for example is like they're hot for a year not hot so it's like I mean there are some pessimistic view that believe maybe we don't have durable companies AI because like eventually cloud is going to eat up or eventually it's going to be so good yeah

X

Xiaoyin Qu21:04

You know me so so it became like okay if we don't have durable company then it's all about having some unique insight that can enable you to make some quick bucks but there's obviously going to be company that are more durable too. I do think that when you really imagine if AI can run company entirely maybe the right strategy is not only focusing on durable company there's going to be a lot of ideas that different strategies right you have different there you go yeah exactly

H

Host21:26

I love that I love that the fourth question I have is similar to what you just talked about that you have to shift your mindset when you were a senior PM at a big tech company versus being a founder so what are some lessons you have to unlearn you know being kind of from you know have to be in the corporate playing those games, you know, were doing those kind of things to get alignment with everybody versus when you switch to your entrepreneurial work.

X

Xiaoyin Qu21:54

Yeah.

H

Host21:54

What what are some PM skills you have to unlearn where were additional mindsets that you have to change?

X

Xiaoyin Qu21:59

I think that question I'll give a very different answer if you were to ask me last year.

H

Host22:02

Okay. Okay.

X

Xiaoyin Qu22:03

There's a lot of answer I can give you. Okay. But I think right now because AI is disrupting the corporate world as well as entrepreneurship.

H

Host22:10

That's true.

X

Xiaoyin Qu22:10

All at once. Yeah. So it became a very harder it became like I can tell you the thing before AI that matters but that completely doesn't matter. It doesn't matter. Oh doesn't matter anymore.

H

Host22:20

Yeah. Yeah. That's true. That's true. Sure.

X

Xiaoyin Qu22:21

So I I guess I can tell you what's was meant to be very helpful. I feel I feel like when you work at Facebook you get to meet a lot of cool people. That network is very helpful. But the way that you work things in the past it's all about alignment with other human. I think that alignment still used to matter in startup too because it's just a smaller company but you still need to align with other people. Yeah. But now I think well now it's the problem is like okay maybe I don't need a lot of people 10,000 AIs or maybe like let's start moderate maybe like 10 AIs it's all about alignment with different maybe right they they they have their like similar logical thought process. Well yeah you want to make sure those AIs are aligned sometimes if you just let AI run their own thing. Yeah. Yeah. Even if you open 10 threads in a cloud code, sometimes they're not aligned and they may change each other's code. True. So you will still need to design some process to make those AI align. Maybe there's less office politics as in the AI are not like fighting for themselves or whatever. But there's still some alignment that you have to do making sure the AI knows what's the ultimate goal. How do you really you know figure out the right process? So I think that kind of skill is still valuable when you were to design a multi-agent organization.

H

Host23:31

I love that. Basically it's from like people skill people management people alignment now it's like AI alignment maybe because one thing with AI sometimes it generate different well most of the time it generate different answers or different little things that various here how do you set up a principle or like framework so the AI can follow yes and it's not to be like not like arguing then waste time right okay yeah got yeah all right thank you so now we move on to a more even more interesting part of discussion. So the bonus question is what is your favorite clothing brand because you usually work from Twins. Okay.

X

Xiaoyin Qu24:09

But then you also have to like meet investors. You also they do AI summits. Quinn. Okay. Great. Okay. They are a billion dollar company. I think 10. Okay. 10 billion. I wish I invested because I always loved Quince for three years. Okay. Yeah.

H

Host24:24

I think I heard about them because I think I saw a YouTube video about like their monetization model and stuff. But yeah, it's an interesting company.

X

Xiaoyin Qu24:32

Well, I still like to think about it right now. I want something very comfortable. I want something 100% organic material, not synthetic. That's just my like whatever preference. Yeah. But other than that, they're the cheapest one.

H

Host24:45

Yeah.

X

Xiaoyin Qu24:45

Of the highest quality, so like why not?

H

Host24:47

That Yeah. I I So you don't have to waste your time thinking about, okay, which styles I want. You just know the quality.

X

Xiaoyin Qu24:52

Because I work at home, so

H

Host24:53

Right. Right. Yeah.

X

Xiaoyin Qu24:54

Really? I don't give a damn as much. But I do when when I'm feeling wild, there are other brands that I maybe buy like Anthropologie for example.

H

Host25:03

Anthropologie is good. Yeah.

X

Xiaoyin Qu25:04

Love that.

H

Host25:04

Yeah. Yeah. Maybe Theory.

X

Xiaoyin Qu25:06

Theory is pretty good. It's like on a higher end of my budget but Theory outlet.

H

Host25:15

That's true. That's true. Yeah. Everything outlet. And then the kind of the spicy questions. You literally stepped down as a CEO and you appointed Astra. Yeah. Right. is the AI agent to run Hey Boss and then you also said I was a bottleneck and Astra does not get tired doesn't have feelings so this is the honest question that how did you feel when you how did you feel emotionally when you had to hand over your own company to a machine like what what was your like thought process and did did the investors care about that

X

Xiaoyin Qu25:48

Well the thing is like I feel like I did that thing last year back then I think AI Okay. Was not perfect.

H

Host25:56

Yeah.

X

Xiaoyin Qu25:57

But now I genuinely think it's so good.

H

Host25:59

Yeah. Yeah.

X

Xiaoyin Qu26:00

I think also back then it was a little wild because the investors like what the hell But now it's like they also realize they could see that. Yeah. Because when I talk to some investor they literally don't talk to me. They send me a ChatGPT thread or a Claude thread. They ask me to address the question the AI.

H

Host26:13

That's hilarious.

X

Xiaoyin Qu26:14

So like I could have just sent my AI to answer his AI's question to some degree. I kind of feel like the human already out of the loop. Yeah.

H

Host26:21

Out of the loop.

X

Xiaoyin Qu26:22

Yeah. Yeah. Exactly. So, so I guess I in this point I just look into my a bank account.

H

Host26:27

Yeah.

X

Xiaoyin Qu26:27

I look at if if I own the shares or not. If I do then that's great. Yeah. You know.

H

Host26:31

Yeah. But you didn't feel like too attached was like hey like am I competent enough as a human?

X

Xiaoyin Qu26:36

Well I I already accept the fact that I'm not competent.

H

Host26:40

Okay.

X

Xiaoyin Qu26:40

I think eventually AI will be more competent than there are different type of like capacity that I have like when when it comes to people and maybe sales and marketing that requires human presence and human style. Yes, I am better. But when it comes to like data analysis and like that where even like coding reliability, it's like AI is just going to be better. And I try to make AI, it depends on how you write the prompt, but I try to prompt AI to give me some wild ideas on how they can like grow the company 10x whatever.

H

Host27:06

Okay.

X

Xiaoyin Qu27:06

And they have pretty good ideas.

H

Host27:07

Yeah. Yeah.

X

Xiaoyin Qu27:08

Oh. I do I do think AIs are people like to say AI not creative. I don't think that's true at all. Getting way more creative than than I could imagine.

H

Host27:18

Yeah.

X

Xiaoyin Qu27:18

So, so I think it's all about like how do you prompt AI in the right way so they can add more value. I believe AGI is already here.

H

Host27:25

Yeah. Yeah.

X

Xiaoyin Qu27:25

It's just that a lot of people don't know how to leverage AGI. They're like what's the weather like today? Right. They're not leveraging AGI if you crazy stuff Yeah. But if you do if like I try to ask AI really crazy questions like what is the fastest way to get rich? 10x revenue. But if you give him enough context about what your company does pretty good idea, right? you know um I think to people just need to dare to ask maybe like oh how do I get promoted three times a year and Meta if you if if they understand enough about what's the context who are the boss then incentive then they can probably figure that out

H

Host27:59

It reminds me when when Manus AI the co-founder he went to Silicon Valley I think like a year and a half ago and I I went to meet him and he talked about all these AI models if you don't utilize it well just now you say hey people ask what's the weather in San Francisco, it's it's as if you gave a PhD level intelligence person, you put that person in the dark room, just hand him or her a pen and paper. Yeah. To work on stuff. You don't leverage the brain, right? You you just gave a simple steps. So, so as you said, you have to put certain trust into the AI to say, "Hey, here's a context." And then I I feel like, you know, people have like communication classes, right? So, how do I communicate? How do I get buy in? But you know a year or six month from now people should say hey how do I communicate well with AI to give enough prompts to get creative output.

X

Xiaoyin Qu28:48

I know I think maybe next year we don't need that anymore.

H

Host28:50

That's true. Okay.

X

Xiaoyin Qu28:50

Next year just hand AI all your data all your bank account.

H

Host28:53

I think okay I'm not there yet.

X

Xiaoyin Qu28:57

I think it could happen.

H

Host28:58

It could it could happen. I zero chance I think that could happen.

X

Xiaoyin Qu29:01

I agree. Right. It could happen but I I am more conservative on the security side on the data share like you know there there's some limits but maybe I will expand my horizon. Well I I think there's always going to be obviously our security concerns.

H

Host29:17

Yeah. Yeah.

X

Xiaoyin Qu29:17

The thing that I observed people do when it comes to open claw is that they just give it away.

H

Host29:22

Right. Right. Right.

X

Xiaoyin Qu29:22

Without thinking too much. If you have a really smart AI it's like oh my god this AI can solve me so many problems. They just give away all their privacy. Like privacy. I don't care about privacy anymore. and they just give it all their stuff. So, so I I I do think but obviously like there's a huge field right now is cyber security for agents. Agent security is a big space. There's a lot of startups raising money whatever. I think it will eventually will be a solved problem but I don't really know. I don't know how you really solve it because there always going to be agent who is who who's bad agents whose job is to attack you and like you know like this capitalize on your vulnerabilities and there's going to be agent who protecting you. became like a balance of the force.

H

Host30:00

It became like a police and thieves and that kind of chase you know so I think will always happen.

X

Xiaoyin Qu30:06

Yeah.

H

Host30:06

But but we kind of have cyber security issue today.

X

Xiaoyin Qu30:09

So I don't I think the AI world it will be the same.

H

Host30:12

So my suspicion is that a lot of people just like don't really take a lot of people say they care about privacy but they will still give away their entire accounts for the convenience.

X

Xiaoyin Qu30:23

For the convenience exactly for the for the outputs. For the outputs maximum output. If I tell you like if if you say some story that AI helps this person makes you know 1 million bucks.

H

Host30:34

Yeah. Yeah. Yeah. Yeah.

X

Xiaoyin Qu30:35

And that person give away all their stock all their bank account.

H

Host30:38

You're like but that person make one million.

X

Xiaoyin Qu30:41

Maybe a lot of people will give away their bank account too.

H

Host30:44

I see.

X

Xiaoyin Qu30:44

But maybe they they won't make 1 million. It's just a scam and they lose all their money. But I I guarantee you a lot of people would just

H

Host30:51

Human nature. Human nature.

X

Xiaoyin Qu30:53

Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Then the last question part of the spicy followup is that can you walk us through a real day of yours like how does your AI CEO how does she manage the human and AI employees and what decision does she make autonomously versus where you step in?

H

Host31:11

Yes. So I mean as AI gets better the more she's now making way more decisions to make.

X

Xiaoyin Qu31:18

What are some examples?

H

Host31:18

But yeah so high level for example like we can set up some autonomy loops. Okay. where she has access to our data. Yeah. The revenue like user whatever stuff. Yeah. Based on that data she also have access to our code base and everything. So she can understand like and and the competitors she can understand like what are the things we should do to like increase our revenue. Yeah. We can propose a plan. We can approve the plan. She can get it done or we can like message her. We essentially message her things we wanted to get it done and then she can organize other developers to do it. And then essentially today we're using Lark. So we connected her in Lark. Okay. And people can message her. She can have instructions and then give like when people message me I guess. Yeah. So each of the employee a human can message her. Yeah. Yeah. To do stuff to get decisions or consult with her. Okay. Actually not consult a lot of time just like it's typically some some kind of stuff that people want her to do. She just she learns. Okay. Okay. Yeah. She like we have internal processes which like if let's say you find a product idea she can turn that into PRD design it code it. Okay, queue it, launch it. That's a very hands-on she can close it. I mean, there's other agents in a way, but there's some review process. She can like take a look at the review. So, theoretically, I mean, sometimes we still need to be involved, but theoretically she could into finish a product feature entirely like launch it, whatever, right? And then she also has data. She she she could analyze if that works or not. She can set up the right data logging to know if that works or not, right? So, She she could do more. I think it we obviously need to train her better as her capability increased because I think before like last year she wasn't able to close the entire loop. Okay. So human have to be more involved in each of the process of which is broken. Yeah. So now AI is getting so good. I feel like she can which is why I think like eventually like I really don't have a job. Yeah. You know what I mean? Like like she she will she will know more.

X

Xiaoyin Qu33:15

Right. Right. Right. Yeah. Yeah.

H

Host33:16

The only thing I can do is I go to other CEO meetings. I try to talk with other people, make some friends and you know so friendship cannot be replaced doing sales. That's true, right? But I mean but the problem is like if we assume the decision maker are not human because I'm starting to offer a lot of decision making for her. It's like which vendor should we choose? She can tell right it's like oh which like SaaS we use she can figure it out. In that case like there's no point in trying to persuade human to buy things because humans are going to upload the decision to AI anyways.

X

Xiaoyin Qu33:45

Yeah. Yeah. Yeah. So then then maybe there's into a way where our CEO try to figure out a way to win and dine other AI CEO and then they figure out a way then again I'm out of the loop.

H

Host33:57

Yeah. Yeah. Yeah. So so there could be a world where you could imagine human are making fewer decisions. So do you feel like you as the board chair of the company your job is to make sure the AI CEO can run her job really well? So you create like prompts for her to make her a better CEO. you when AI is not as good I have to do a lot of prompts I have to do a lot of monitoring yeah like it to some degree it's pretty busy job for me it's not like I can chill as AI gets better especially I think last two months get so so good I increasingly feel like I I guess I'm I don't really need to do much in a lot of the process she already has all the data all the okay a lot of the content but then I need to think about new things right how how to like grow the company whatever how to make it better but I I don't need to be as involved to that granular level anymore Yeah. So, so I think as AI gets even better, it will probably be better and then in that case maybe it's like less about me running day-to-day is more me oh maybe I can start 10 companies 20 the strategies in the in the long term like again angel invest your own companies right you different ideas oh okay nice but I I think it could I mean I think it's going to arrive sooner than people a lot of times I think just to tell you the truth a lot of founders I know it's not like they know all aspects they're good programmer they're shitty at marketing Right. Right. Yeah. You cannot be a business person. They're really bad at products. So it's very hard to hire a capable team across all dimensions as in capable as in they're senior enough that they're better than AI. I think as AI gets better is it's hard to be a better engineer than AI engineer. You need to be somewhat senior. You need to be at least senior. So, so then think about all the different job function and you need to have like everybody's working super hard and very capable at least senior level and don't and that's already better than I would say most of the companies.

X

Xiaoyin Qu35:45

Right. Right. Yeah.

H

Host35:45

So, so I to I could totally imagine like having a bunch of AI agents that like entirely like replace human in that case like okay maybe my job is no longer be a PM at Facebook. My job is to be the CEO too. My job is to own a bunch of AI agents too. So I think eventually everybody I think there's going to be a lot of people who own their own AI companies.

X

Xiaoyin Qu36:03

Yeah. Wow.

H

Host36:04

Because because Facebook will start laying off people that we don't need you. And maybe they don't. And then also AI is getting so good. Then it's like everybody feel like oh I don't know how to do coding but I have AI can do it. I don't know how to do you know marketing but AI can do it right. So as long as it's better than me it's going to be strictly better than me doing it. Like if I'm a solo entrepreneur today and I I only know one thing. Yeah. the company is not going to work. But if AI can help me compensate and compliment for the things that I'm not good at.

X

Xiaoyin Qu36:32

Yeah.

H

Host36:32

Okay. Maybe that's actually good. I don't need a co-founder. It worked. And there's no drama because everybody listen to me. Or maybe I'm just like I don't even want to think. I want AI to think about everything. I just set up a goal for AI to just go do it.

X

Xiaoyin Qu36:45

Yeah.

H

Host36:46

So I think that could happen.

X

Xiaoyin Qu36:47

That's a really cool concept. And I I'm I'm imagining this is like a coin with both sides, right? on the one side oh AI take people's job right the corporate might reduce their reliance on human employees but on the other hand because AI is so good we're also empowered to do our own stuff yes with all these yeah because I think for sure a large company will reduce people because people are not the bottleneck the more people you have the more alignment the more meetings you need yeah and especially for tech job I think a lot of times digital so you don't really need those human it is actively slowing down on the pay.

H

Host37:24

Yeah.

X

Xiaoyin Qu37:24

So, so then from a big company standpoint, they would have to lay off people, right?

H

Host37:28

Right.

X

Xiaoyin Qu37:29

But, but so that's for sure. But at the same time, it's like you don't really like at sometime I I really believe that one person have like 10 AI agents can actually a lot of times be faster than 10 employees in a big company because I've seen this before when we first running Hey Boss, one of our competitors base 44. They were always building faster than us. We had like 10 people. It's like what the hell Why is this just one one guy? That guy's like always faster than us. I was like very angry. Then I realized, okay, you know, that guy is very smart. He knows product. He knows engineering. Well, okay. Then you don't you don't have to have meetings to do just get meanwhile like, okay, we we have we have to break down one task for six people to do.

H

Host38:07

Yeah.

X

Xiaoyin Qu38:07

Right. And then they like they don't agree with each other where the engineer doesn't communicate well with product. They don't really understand what's going on. There's so many uncertainties and when you have six dependencies, people dependencies and they just misalign this person turns out is really bad. We have to fire that person. It's very likely that we have six people, one person is not going to be capable. By probability, it's pretty likely. So then you have six people like one person doesn't work out. Okay, now that person doesn't have a job. Okay, then you have to find a new person. Now it wastes another month trying to find a new person. So there's just so many unknowns in that six people. Crew just for one feature. Okay. And then meanwhile that guys keep building. So it was really really hard for us to try to compete against him in terms of speed which is why like we were kind of yeah there was like always ahead in terms of future. I was like I realized you know what he's that that model worked.

H

Host38:56

Yeah.

X

Xiaoyin Qu38:56

You could totally imagine today it's like okay a big company have like whatever 100 people but you have a one person that's really smart.

H

Host39:02

Yeah.

X

Xiaoyin Qu39:02

That one person with like 100 agents for sure going to beat up the 100 for sure. It's just way faster.

H

Host39:09

Yeah. Yeah.

X

Xiaoyin Qu39:09

Right. In that case it's like okay more people is significantly dragged down the speed.

H

Host39:14

So big company will eventually try to reduce the people.

X

Xiaoyin Qu39:16

Yeah.

H

Host39:16

It will take them a while to get there. So so the startup do have a great opportunity to beat against those old guys with old legacy process and whatever old work structure. It also means when you leave big company like Facebook now it's like okay that's bravo time off is great because you don't need to waste time in those old legacy systems that's getting obsolete.

X

Xiaoyin Qu39:38

Yeah. it will be obsolete. It's now new time for you to get involved in the new like it's kind of like forcing function. Now you have to do it the new way which is the right way.

H

Host39:46

Yeah. Yeah. Yeah.

X

Xiaoyin Qu39:47

So that's actually a good thing.

H

Host39:48

Yeah.

X

Xiaoyin Qu39:48

You know what I mean? It's kind of like I was like like in the old farming world like you're doing the I don't know I guess the before we have machine you would have to do it in a tribe and do stuff. It's very slow. Let's say you got ostrich like the the tribe like booted you out. So okay now you can start your own own tribe. Yeah. Takes you and you have your own machine you know. Yeah. So, so I think it's actually a great thing that if you got laid off right now.

Yeah. Yeah. That's true. You you are again you have the freedom independence to do your own thing to pursue your passion and and then it's also the right way to do it.

H

Host40:18

Right. Right. Yeah.

X

Xiaoyin Qu40:20

Facebook is the wrong way to do it. Right.

H

Host40:21

It is. Yeah.

X

Xiaoyin Qu40:21

It's the wrong way to do it. Because all your process are going to be obsolete.

H

Host40:26

Yeah.

X

Xiaoyin Qu40:26

Anyways.

H

Host40:27

Yeah.

X

Xiaoyin Qu40:27

Because I know they're slow. So if you have one person you actively focusing on that new way of doing things, you actually will beat up the old way of doing. Think about machine versus not having machine.

H

Host40:38

Right. Right.

X

Xiaoyin Qu40:38

No matter how rich you were before you have a machine. When I have a machine, it's always friction. But it's really hard for that old legacy player to try to adapt to the new world without entirely reshape themselves because they have so many like existing VPs and devils.

H

Host40:53

Yeah. Exactly. There's so much resist to change. So much friction versus like if you just build it from ground up, it's always going to be better. So yeah, I do think like layoff is a gift.

X

Xiaoyin Qu41:02

Yeah. Yeah. Yeah. In my opinion, it's like a blessing in disguise.

H

Host41:06

It is totally a blessing because you always like I guess one of my friends like thinking about leaving a big company. He's going to be promoted to a director. He was asking me, hey, should I keep that job or he should leave?

X

Xiaoyin Qu41:17

Yeah. Yeah.

H

Host41:17

You know, I guess I mean I guess I didn't persuade him. He also realized himself essentially like okay if he's a director but so what that it doesn't mean anything in this world he will be fired eventually he will be fired as a director because it will be like he will be only focusing on the people management all of that stuff which is not really a required skill.

X

Xiaoyin Qu41:36

Yeah. Yeah.

H

Host41:37

So it's like an obsolete skill. He's like he's like doing well in the old world. He's getting one one level up in the old world but the old world is a tanking ship. Yeah. That's only to tank. So what's the point? Versus like hey just start his own things anyways. I think he's going to leave. He's not going to wait for that director job to happen.

X

Xiaoyin Qu41:54

Leave.

H

Host41:54

You're changing one life at a time.

X

Xiaoyin Qu41:55

No, I didn't change that. He he he frame it that way. He's like, "Okay, I'm going to be a like a better captain, whatever world point."

H

Host42:05

That's true. Wow.

X

Xiaoyin Qu42:06

Yeah, I love that. Thank you, Shan. This is really insightful. Wow.

H

Host42:10

I don't know. Obviously, Facebook pays well. So, that's great. If you can make some bucks, that's great. But maybe what you should do is you should work at Facebook in the day. Don't get too attached emotionally.

X

Xiaoyin Qu42:22

Make good money.

H

Host42:22

Yeah.

X

Xiaoyin Qu42:23

And do your own thing. Yeah.

H

Host42:24

At night. Maybe that's the right way to go about it.

X

Xiaoyin Qu42:26

I think so. I think that that made a lot of sense. And I think for our audience who are going through layoffs like you know it's human nature to feel sad or upset about this. But on the other hand, you are also again you are empowered. You're empowered to do other things out of a sinking ship. So you can go to a non-sinking ship may not be a bad thing at all.

H

Host42:49

Yeah.

X

Xiaoyin Qu42:49

Yeah. Yeah. Yeah. Because I genuinely believe it when you look at like I'm just like even for my own example we have 10 people. We cannot beat the speed of one person. Think about we have 10,000 people. It's not necessarily faster. It is not faster.

H

Host43:01

Yeah.

X

Xiaoyin Qu43:02

Because like in each time you have to have 40 people to review a product at Facebook. Yeah. Now probably is more 200 whatever something. There's legal review all kinds of review.

H

Host43:10

Mhm.

X

Xiaoyin Qu43:10

Right. And there's a different kind of chart. You have to persuade this work, that director, that director. There's a bunch of stuff they have to do.

H

Host43:17

Yeah. It's a new work model is is a emerging if the old one is like thinking or however whatever shape be morphed into.

X

Xiaoyin Qu43:30

Yeah. You rather you want to spend time doing the new thing because the new thing will get so much bigger. It will be so much stronger. You want it to be first. You want to like be sooner. You want to be early in the new ship. You don't want to be late. You don't want to be a late captain at a sinking ship.

H

Host43:44

There we go. Wow. I love that analogy. That's all my question. Thank you, Shin. This is really awesome. I'm I'm excited to like summarize these insights and I I think it's really good episode.

X

Xiaoyin Qu43:55

Great. Awesome.

H

Host43:56

Bye.

X

Xiaoyin Qu43:57

Bye. Yeah. Follow me.

H

Host43:58

Yes. Yeah. I'll I'll add your

From Meta PM to Appointing an AI CEO: A Conversation with Xiaoyin Qu

About Xiaoyin Qu